{
    "info": {
        "author": "yliess",
        "author_email": "hatiyliess86@gmail.com",
        "bugtrack_url": null,
        "classifiers": [],
        "description": "\nTriplet Loss Utility for Pytorch Library.\n\n# TripletTorch\n\nTripletTorch is a small pytorch utility for triplet loss projects. It provides\nsimple way to create custom triplet datasets and common triplet mining loss\ntechniques.\n\n## Install\n\nInstall the module using the pip utility ( may require to run as sudo ).\n\n```bash\npip3 install triplettorch\n```\n\n## Usage\n\n### Triplet Dataset\n\n```python\nfrom triplettorch import TripletDataset\n\n# Create a triplet dataset given:\n#   * labels  : array of label ( class ) for each sample of the dataset\n#   * data_fn : method to access data for a given index in the dataset\n#   * size    : number of samples in the dataset\n#   * n_sample: number of sample per draw ( to increase probability to\n#               contain valid triplets in a batch )\n# Do not forget to concatenate batch dimension and sample dimension\n# when used with a DataLoader as TripletDataset[ idx ] returns a\n# ( batch_size, n_sample, ... ) dimension tensor for labels and data\ndataset = TripletDataset( labels, data_fn, size, n_sample )\n```\n\n### Triplet Mining\n\n```python\nfrom triplettorch import AllTripletMiner, HardNegativeTripletMiner\n\n# Define the triplet mining loss given:\n#   * margin: the margin float value from the triplet loss definition\nminer          = AllTripletMiner( .5 ).cuda( )\nminer          = HardNegativeTripletMiner( .5 ).cuda( )\n\n# Use the loss in training given:\n#   * labels    : array of label ( class ) for each sample of the batch\n#   * embeddings: output of the neural network for each sample of the batch\n# Returns two values:\n#   * loss    : triplet loss value\n#   * frac_pos: fraction of positive triplets\n#               None ( None HardNegativeTripletMiner )\nloss, frac_pos = miner( labels, embeddings )\n```\n\n## Example\n\nThe repository provides an example application with the MNIST dataset.\n\n![ MNIST ]( examples/MNIST_AllTripletMiner.png )\n\n\n## References\n* [FaceNet: A Unified Embedding for Face Recognition and Clustering]\n* [Triplet Loss and Online Triplet Mining in TensorFlow]\n\n[FaceNet: A Unified Embedding for Face Recognition and Clustering]: https://arxiv.org/pdf/1503.03832.pdf\n[Triplet Loss and Online Triplet Mining in TensorFlow]:https://omoindrot.github.io/triplet-loss\n\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/TowardHumanizedInteraction/TripletTorch",
        "keywords": "",
        "license": "MIT",
        "maintainer": "",
        "maintainer_email": "",
        "name": "TripletTorch",
        "package_url": "https://pypi.org/project/TripletTorch/",
        "platform": "",
        "project_url": "https://pypi.org/project/TripletTorch/",
        "project_urls": {
            "Homepage": "https://github.com/TowardHumanizedInteraction/TripletTorch"
        },
        "release_url": "https://pypi.org/project/TripletTorch/0.1.2/",
        "requires_dist": [
            "numpy (==1.17.2)",
            "torch (==1.2.0)",
            "torchvision (==0.4.0)",
            "tqdm (==4.36.1)",
            "matplotlib (==3.1.1)"
        ],
        "requires_python": ">=3.4, <4",
        "summary": "Triplet Loss Utils for Pytorch Library.",
        "version": "0.1.2"
    },
    "last_serial": 5936256,
    "releases": {
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "0cbb69efa3bc9631f1b532b25f149a61",
                    "sha256": "1e917cc54e6d097c72e523dd6fd6a88fa6f92c6bd8e724895ca49063346a11c0"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.0-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "0cbb69efa3bc9631f1b532b25f149a61",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">=3.4, <4",
                "size": 5228,
                "upload_time": "2019-10-06T21:49:31",
                "url": "https://files.pythonhosted.org/packages/de/c6/742aeb3951abdaf6e192341e3eb8dce168369befa196106f8ba3b093a434/TripletTorch-0.1.0-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "fe73745d2b3ab95d455f29efb922f194",
                    "sha256": "e2334d8428f7f06d133efd5cb01756033e77aad8774d4570c4f3d95de2564247"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "fe73745d2b3ab95d455f29efb922f194",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4, <4",
                "size": 4484,
                "upload_time": "2019-10-06T21:49:34",
                "url": "https://files.pythonhosted.org/packages/06/72/665194e3fd42331f040d9e26a70562dcd48ff232bdb507626e4eadea5f94/TripletTorch-0.1.0.tar.gz"
            }
        ],
        "0.1.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "7651e31a897763b3632238ca47a73188",
                    "sha256": "4dab7890c209d327b2ca84ea61ffeb0c919c796e9248db0ec6041bd04aa4b529"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "7651e31a897763b3632238ca47a73188",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">=3.4, <4",
                "size": 5259,
                "upload_time": "2019-10-06T21:52:25",
                "url": "https://files.pythonhosted.org/packages/c2/31/673bc2686b6a61dd16642c7eedc59e493d899130e0e5d9ce6875754175dd/TripletTorch-0.1.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "3e4f6f5ef3eae74b355faef940b09641",
                    "sha256": "530a0d6303afab8c42442761bf60e03b1f3ccb1d896087b93902ce2509a5c19a"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.1.tar.gz",
                "has_sig": false,
                "md5_digest": "3e4f6f5ef3eae74b355faef940b09641",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4, <4",
                "size": 4486,
                "upload_time": "2019-10-06T21:52:27",
                "url": "https://files.pythonhosted.org/packages/d6/5c/5987b4989fda6759ad00e825206a0e50dbc01e3840064fe313de6e4d915f/TripletTorch-0.1.1.tar.gz"
            }
        ],
        "0.1.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "5424f3076bc867253b923c65e1bb110b",
                    "sha256": "1825094af01e067d7afc2eed46fa58e3ceb51ff9f32febe31486152be9d8e991"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.2-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "5424f3076bc867253b923c65e1bb110b",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">=3.4, <4",
                "size": 6130,
                "upload_time": "2019-10-06T22:53:41",
                "url": "https://files.pythonhosted.org/packages/07/d3/1c9f8221e173ed491c604472876e533beccb62ca2e20fd841e8c7d78b203/TripletTorch-0.1.2-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "61357944c38f8321cc52a92fce184941",
                    "sha256": "f8a73c3eae348e3a56c2fafe3d7b20d962392b77604ee6ba3a9691eff87d7fd6"
                },
                "downloads": -1,
                "filename": "TripletTorch-0.1.2.tar.gz",
                "has_sig": false,
                "md5_digest": "61357944c38f8321cc52a92fce184941",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4, <4",
                "size": 5502,
                "upload_time": "2019-10-06T22:53:42",
                "url": "https://files.pythonhosted.org/packages/09/42/53d62e2c287baf452e8d3f05814dfaaf745e76e20e13582cde77ccc08bd5/TripletTorch-0.1.2.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "5424f3076bc867253b923c65e1bb110b",
                "sha256": "1825094af01e067d7afc2eed46fa58e3ceb51ff9f32febe31486152be9d8e991"
            },
            "downloads": -1,
            "filename": "TripletTorch-0.1.2-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "5424f3076bc867253b923c65e1bb110b",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": ">=3.4, <4",
            "size": 6130,
            "upload_time": "2019-10-06T22:53:41",
            "url": "https://files.pythonhosted.org/packages/07/d3/1c9f8221e173ed491c604472876e533beccb62ca2e20fd841e8c7d78b203/TripletTorch-0.1.2-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "61357944c38f8321cc52a92fce184941",
                "sha256": "f8a73c3eae348e3a56c2fafe3d7b20d962392b77604ee6ba3a9691eff87d7fd6"
            },
            "downloads": -1,
            "filename": "TripletTorch-0.1.2.tar.gz",
            "has_sig": false,
            "md5_digest": "61357944c38f8321cc52a92fce184941",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.4, <4",
            "size": 5502,
            "upload_time": "2019-10-06T22:53:42",
            "url": "https://files.pythonhosted.org/packages/09/42/53d62e2c287baf452e8d3f05814dfaaf745e76e20e13582cde77ccc08bd5/TripletTorch-0.1.2.tar.gz"
        }
    ]
}